target: encapsulate smp_mb__after_atomic()

The target code has a rather generous helping of smp_mb__after_atomic()
throughout the code base.  Most atomic operations were followed by one
and none were preceded by smp_mb__before_atomic(), nor accompanied by a
comment explaining the need for a barrier.

Instead of trying to prove for every case whether or not it is needed,
this patch introduces atomic_inc_mb() and atomic_dec_mb(), which
explicitly include the memory barriers before and after the atomic
operation.  For now they are defined in a target header, although they
could be of general use.

Most of the existing atomic/mb combinations were replaced by the new
helpers.  In a few cases the atomic was sandwiched in
spin_lock/spin_unlock and I simply removed the barrier.

I suspect that in most cases the correct conversion would have been to
drop the barrier.  I also suspect that a few cases exist where a) the
barrier was necessary and b) a second barrier before the atomic would
have been necessary and got added by this patch.

Signed-off-by: Joern Engel <joern@logfs.org>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>

6 files changed, 56 insertions, 115 deletions

diff --git a/drivers/target/loopback/tcm_loop.c b/drivers/target/loopback/tcm_loop.c
index 340de9d92b15..a7f6dc646045 100644
--- a/drivers/target/loopback/tcm_loop.c
+++ b/drivers/target/loopback/tcm_loop.c
@@ -960,8 +960,7 @@ static int tcm_loop_port_link(
 				struct tcm_loop_tpg, tl_se_tpg);
 	struct tcm_loop_hba *tl_hba = tl_tpg->tl_hba;
 
-	atomic_inc(&tl_tpg->tl_tpg_port_count);
-	smp_mb__after_atomic();
+	atomic_inc_mb(&tl_tpg->tl_tpg_port_count);
 	/*
 	 * Add Linux/SCSI struct scsi_device by HCTL
 	 */
@@ -995,8 +994,7 @@ static void tcm_loop_port_unlink(
 	scsi_remove_device(sd);
 	scsi_device_put(sd);
 
-	atomic_dec(&tl_tpg->tl_tpg_port_count);
-	smp_mb__after_atomic();
+	atomic_dec_mb(&tl_tpg->tl_tpg_port_count);
 
 	pr_debug("TCM_Loop_ConfigFS: Port Unlink Successful\n");
 }
diff --git a/drivers/target/target_core_alua.c b/drivers/target/target_core_alua.c
index fbc5ebb5f761..fb87780929d2 100644
--- a/drivers/target/target_core_alua.c
+++ b/drivers/target/target_core_alua.c
@@ -392,8 +392,7 @@ target_emulate_set_target_port_groups(struct se_cmd *cmd)
 				if (tg_pt_id != tg_pt_gp->tg_pt_gp_id)
 					continue;
 
-				atomic_inc(&tg_pt_gp->tg_pt_gp_ref_cnt);
-				smp_mb__after_atomic();
+				atomic_inc_mb(&tg_pt_gp->tg_pt_gp_ref_cnt);
 
 				spin_unlock(&dev->t10_alua.tg_pt_gps_lock);
 
@@ -403,8 +402,7 @@ target_emulate_set_target_port_groups(struct se_cmd *cmd)
 					found = true;
 
 				spin_lock(&dev->t10_alua.tg_pt_gps_lock);
-				atomic_dec(&tg_pt_gp->tg_pt_gp_ref_cnt);
-				smp_mb__after_atomic();
+				atomic_dec_mb(&tg_pt_gp->tg_pt_gp_ref_cnt);
 				break;
 			}
 			spin_unlock(&dev->t10_alua.tg_pt_gps_lock);
@@ -998,8 +996,7 @@ static void core_alua_do_transition_tg_pt_work(struct work_struct *work)
 		 * every I_T nexus other than the I_T nexus on which the SET
 		 * TARGET PORT GROUPS command
 		 */
-		atomic_inc(&mem->tg_pt_gp_mem_ref_cnt);
-		smp_mb__after_atomic();
+		atomic_inc_mb(&mem->tg_pt_gp_mem_ref_cnt);
 		spin_unlock(&tg_pt_gp->tg_pt_gp_lock);
 
 		spin_lock_bh(&port->sep_alua_lock);
@@ -1028,8 +1025,7 @@ static void core_alua_do_transition_tg_pt_work(struct work_struct *work)
 		spin_unlock_bh(&port->sep_alua_lock);
 
 		spin_lock(&tg_pt_gp->tg_pt_gp_lock);
-		atomic_dec(&mem->tg_pt_gp_mem_ref_cnt);
-		smp_mb__after_atomic();
+		atomic_dec_mb(&mem->tg_pt_gp_mem_ref_cnt);
 	}
 	spin_unlock(&tg_pt_gp->tg_pt_gp_lock);
 	/*
@@ -1063,7 +1059,6 @@ static void core_alua_do_transition_tg_pt_work(struct work_struct *work)
 		core_alua_dump_state(tg_pt_gp->tg_pt_gp_alua_pending_state));
 	spin_lock(&dev->t10_alua.tg_pt_gps_lock);
 	atomic_dec(&tg_pt_gp->tg_pt_gp_ref_cnt);
-	smp_mb__after_atomic();
 	spin_unlock(&dev->t10_alua.tg_pt_gps_lock);
 
 	if (tg_pt_gp->tg_pt_gp_transition_complete)
@@ -1125,7 +1120,6 @@ static int core_alua_do_transition_tg_pt(
 	 */
 	spin_lock(&dev->t10_alua.tg_pt_gps_lock);
 	atomic_inc(&tg_pt_gp->tg_pt_gp_ref_cnt);
-	smp_mb__after_atomic();
 	spin_unlock(&dev->t10_alua.tg_pt_gps_lock);
 
 	if (!explicit && tg_pt_gp->tg_pt_gp_implicit_trans_secs) {
@@ -1168,7 +1162,6 @@ int core_alua_do_port_transition(
 	spin_lock(&local_lu_gp_mem->lu_gp_mem_lock);
 	lu_gp = local_lu_gp_mem->lu_gp;
 	atomic_inc(&lu_gp->lu_gp_ref_cnt);
-	smp_mb__after_atomic();
 	spin_unlock(&local_lu_gp_mem->lu_gp_mem_lock);
 	/*
 	 * For storage objects that are members of the 'default_lu_gp',
@@ -1184,8 +1177,7 @@ int core_alua_do_port_transition(
 		l_tg_pt_gp->tg_pt_gp_alua_nacl = l_nacl;
 		rc = core_alua_do_transition_tg_pt(l_tg_pt_gp,
 						   new_state, explicit);
-		atomic_dec(&lu_gp->lu_gp_ref_cnt);
-		smp_mb__after_atomic();
+		atomic_dec_mb(&lu_gp->lu_gp_ref_cnt);
 		return rc;
 	}
 	/*
@@ -1198,8 +1190,7 @@ int core_alua_do_port_transition(
 				lu_gp_mem_list) {
 
 		dev = lu_gp_mem->lu_gp_mem_dev;
-		atomic_inc(&lu_gp_mem->lu_gp_mem_ref_cnt);
-		smp_mb__after_atomic();
+		atomic_inc_mb(&lu_gp_mem->lu_gp_mem_ref_cnt);
 		spin_unlock(&lu_gp->lu_gp_lock);
 
 		spin_lock(&dev->t10_alua.tg_pt_gps_lock);
@@ -1227,8 +1218,7 @@ int core_alua_do_port_transition(
 				tg_pt_gp->tg_pt_gp_alua_port = NULL;
 				tg_pt_gp->tg_pt_gp_alua_nacl = NULL;
 			}
-			atomic_inc(&tg_pt_gp->tg_pt_gp_ref_cnt);
-			smp_mb__after_atomic();
+			atomic_inc_mb(&tg_pt_gp->tg_pt_gp_ref_cnt);
 			spin_unlock(&dev->t10_alua.tg_pt_gps_lock);
 			/*
 			 * core_alua_do_transition_tg_pt() will always return
@@ -1238,16 +1228,14 @@ int core_alua_do_port_transition(
 					new_state, explicit);
 
 			spin_lock(&dev->t10_alua.tg_pt_gps_lock);
-			atomic_dec(&tg_pt_gp->tg_pt_gp_ref_cnt);
-			smp_mb__after_atomic();
+			atomic_dec_mb(&tg_pt_gp->tg_pt_gp_ref_cnt);
 			if (rc)
 				break;
 		}
 		spin_unlock(&dev->t10_alua.tg_pt_gps_lock);
 
 		spin_lock(&lu_gp->lu_gp_lock);
-		atomic_dec(&lu_gp_mem->lu_gp_mem_ref_cnt);
-		smp_mb__after_atomic();
+		atomic_dec_mb(&lu_gp_mem->lu_gp_mem_ref_cnt);
 	}
 	spin_unlock(&lu_gp->lu_gp_lock);
 
@@ -1260,8 +1248,7 @@ int core_alua_do_port_transition(
 			 core_alua_dump_state(new_state));
 	}
 
-	atomic_dec(&lu_gp->lu_gp_ref_cnt);
-	smp_mb__after_atomic();
+	atomic_dec_mb(&lu_gp->lu_gp_ref_cnt);
 	return rc;
 }
 
diff --git a/drivers/target/target_core_device.c b/drivers/target/target_core_device.c
index e784284cbc2e..f5057a2f4ed1 100644
--- a/drivers/target/target_core_device.c
+++ b/drivers/target/target_core_device.c
@@ -224,8 +224,7 @@ struct se_dev_entry *core_get_se_deve_from_rtpi(
 		if (port->sep_rtpi != rtpi)
 			continue;
 
-		atomic_inc(&deve->pr_ref_count);
-		smp_mb__after_atomic();
+		atomic_inc_mb(&deve->pr_ref_count);
 		spin_unlock_irq(&nacl->device_list_lock);
 
 		return deve;
@@ -1388,8 +1387,7 @@ int core_dev_add_initiator_node_lun_acl(
 
 	spin_lock(&lun->lun_acl_lock);
 	list_add_tail(&lacl->lacl_list, &lun->lun_acl_list);
-	atomic_inc(&lun->lun_acl_count);
-	smp_mb__after_atomic();
+	atomic_inc_mb(&lun->lun_acl_count);
 	spin_unlock(&lun->lun_acl_lock);
 
 	pr_debug("%s_TPG[%hu]_LUN[%u->%u] - Added %s ACL for "
@@ -1422,8 +1420,7 @@ int core_dev_del_initiator_node_lun_acl(
 
 	spin_lock(&lun->lun_acl_lock);
 	list_del(&lacl->lacl_list);
-	atomic_dec(&lun->lun_acl_count);
-	smp_mb__after_atomic();
+	atomic_dec_mb(&lun->lun_acl_count);
 	spin_unlock(&lun->lun_acl_lock);
 
 	core_disable_device_list_for_node(lun, NULL, lacl->mapped_lun,
diff --git a/drivers/target/target_core_pr.c b/drivers/target/target_core_pr.c
index 281d52e3fe99..48a801045176 100644
--- a/drivers/target/target_core_pr.c
+++ b/drivers/target/target_core_pr.c
@@ -674,8 +674,7 @@ static struct t10_pr_registration *__core_scsi3_alloc_registration(
 	 */
 	spin_lock(&dev->se_port_lock);
 	list_for_each_entry_safe(port, port_tmp, &dev->dev_sep_list, sep_list) {
-		atomic_inc(&port->sep_tg_pt_ref_cnt);
-		smp_mb__after_atomic();
+		atomic_inc_mb(&port->sep_tg_pt_ref_cnt);
 		spin_unlock(&dev->se_port_lock);
 
 		spin_lock_bh(&port->sep_alua_lock);
@@ -709,8 +708,7 @@ static struct t10_pr_registration *__core_scsi3_alloc_registration(
 			if (strcmp(nacl->initiatorname, nacl_tmp->initiatorname))
 				continue;
 
-			atomic_inc(&deve_tmp->pr_ref_count);
-			smp_mb__after_atomic();
+			atomic_inc_mb(&deve_tmp->pr_ref_count);
 			spin_unlock_bh(&port->sep_alua_lock);
 			/*
 			 * Grab a configfs group dependency that is released
@@ -722,10 +720,8 @@ static struct t10_pr_registration *__core_scsi3_alloc_registration(
 			if (ret < 0) {
 				pr_err("core_scsi3_lunacl_depend"
 						"_item() failed\n");
-				atomic_dec(&port->sep_tg_pt_ref_cnt);
-				smp_mb__after_atomic();
-				atomic_dec(&deve_tmp->pr_ref_count);
-				smp_mb__after_atomic();
+				atomic_dec_mb(&port->sep_tg_pt_ref_cnt);
+				atomic_dec_mb(&deve_tmp->pr_ref_count);
 				goto out;
 			}
 			/*
@@ -739,10 +735,8 @@ static struct t10_pr_registration *__core_scsi3_alloc_registration(
 						nacl_tmp, deve_tmp, NULL,
 						sa_res_key, all_tg_pt, aptpl);
 			if (!pr_reg_atp) {
-				atomic_dec(&port->sep_tg_pt_ref_cnt);
-				smp_mb__after_atomic();
-				atomic_dec(&deve_tmp->pr_ref_count);
-				smp_mb__after_atomic();
+				atomic_dec_mb(&port->sep_tg_pt_ref_cnt);
+				atomic_dec_mb(&deve_tmp->pr_ref_count);
 				core_scsi3_lunacl_undepend_item(deve_tmp);
 				goto out;
 			}
@@ -754,8 +748,7 @@ static struct t10_pr_registration *__core_scsi3_alloc_registration(
 		spin_unlock_bh(&port->sep_alua_lock);
 
 		spin_lock(&dev->se_port_lock);
-		atomic_dec(&port->sep_tg_pt_ref_cnt);
-		smp_mb__after_atomic();
+		atomic_dec_mb(&port->sep_tg_pt_ref_cnt);
 	}
 	spin_unlock(&dev->se_port_lock);
 
@@ -1109,8 +1102,7 @@ static struct t10_pr_registration *__core_scsi3_locate_pr_reg(
 				if (dev->dev_attrib.enforce_pr_isids)
 					continue;
 			}
-			atomic_inc(&pr_reg->pr_res_holders);
-			smp_mb__after_atomic();
+			atomic_inc_mb(&pr_reg->pr_res_holders);
 			spin_unlock(&pr_tmpl->registration_lock);
 			return pr_reg;
 		}
@@ -1124,8 +1116,7 @@ static struct t10_pr_registration *__core_scsi3_locate_pr_reg(
 		if (strcmp(isid, pr_reg->pr_reg_isid))
 			continue;
 
-		atomic_inc(&pr_reg->pr_res_holders);
-		smp_mb__after_atomic();
+		atomic_inc_mb(&pr_reg->pr_res_holders);
 		spin_unlock(&pr_tmpl->registration_lock);
 		return pr_reg;
 	}
@@ -1154,8 +1145,7 @@ static struct t10_pr_registration *core_scsi3_locate_pr_reg(
 
 static void core_scsi3_put_pr_reg(struct t10_pr_registration *pr_reg)
 {
-	atomic_dec(&pr_reg->pr_res_holders);
-	smp_mb__after_atomic();
+	atomic_dec_mb(&pr_reg->pr_res_holders);
 }
 
 static int core_scsi3_check_implicit_release(
@@ -1348,8 +1338,7 @@ static void core_scsi3_tpg_undepend_item(struct se_portal_group *tpg)
 	configfs_undepend_item(tpg->se_tpg_tfo->tf_subsys,
 			&tpg->tpg_group.cg_item);
 
-	atomic_dec(&tpg->tpg_pr_ref_count);
-	smp_mb__after_atomic();
+	atomic_dec_mb(&tpg->tpg_pr_ref_count);
 }
 
 static int core_scsi3_nodeacl_depend_item(struct se_node_acl *nacl)
@@ -1368,16 +1357,14 @@ static void core_scsi3_nodeacl_undepend_item(struct se_node_acl *nacl)
 	struct se_portal_group *tpg = nacl->se_tpg;
 
 	if (nacl->dynamic_node_acl) {
-		atomic_dec(&nacl->acl_pr_ref_count);
-		smp_mb__after_atomic();
+		atomic_dec_mb(&nacl->acl_pr_ref_count);
 		return;
 	}
 
 	configfs_undepend_item(tpg->se_tpg_tfo->tf_subsys,
 			&nacl->acl_group.cg_item);
 
-	atomic_dec(&nacl->acl_pr_ref_count);
-	smp_mb__after_atomic();
+	atomic_dec_mb(&nacl->acl_pr_ref_count);
 }
 
 static int core_scsi3_lunacl_depend_item(struct se_dev_entry *se_deve)
@@ -1407,8 +1394,7 @@ static void core_scsi3_lunacl_undepend_item(struct se_dev_entry *se_deve)
 	 * For nacl->dynamic_node_acl=1
 	 */
 	if (!lun_acl) {
-		atomic_dec(&se_deve->pr_ref_count);
-		smp_mb__after_atomic();
+		atomic_dec_mb(&se_deve->pr_ref_count);
 		return;
 	}
 	nacl = lun_acl->se_lun_nacl;
@@ -1417,8 +1403,7 @@ static void core_scsi3_lunacl_undepend_item(struct se_dev_entry *se_deve)
 	configfs_undepend_item(tpg->se_tpg_tfo->tf_subsys,
 			&lun_acl->se_lun_group.cg_item);
 
-	atomic_dec(&se_deve->pr_ref_count);
-	smp_mb__after_atomic();
+	atomic_dec_mb(&se_deve->pr_ref_count);
 }
 
 static sense_reason_t
@@ -1551,15 +1536,13 @@ core_scsi3_decode_spec_i_port(
 			if (!i_str)
 				continue;
 
-			atomic_inc(&tmp_tpg->tpg_pr_ref_count);
-			smp_mb__after_atomic();
+			atomic_inc_mb(&tmp_tpg->tpg_pr_ref_count);
 			spin_unlock(&dev->se_port_lock);
 
 			if (core_scsi3_tpg_depend_item(tmp_tpg)) {
 				pr_err(" core_scsi3_tpg_depend_item()"
 					" for tmp_tpg\n");
-				atomic_dec(&tmp_tpg->tpg_pr_ref_count);
-				smp_mb__after_atomic();
+				atomic_dec_mb(&tmp_tpg->tpg_pr_ref_count);
 				ret = TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE;
 				goto out_unmap;
 			}
@@ -1571,10 +1554,8 @@ core_scsi3_decode_spec_i_port(
 			spin_lock_irq(&tmp_tpg->acl_node_lock);
 			dest_node_acl = __core_tpg_get_initiator_node_acl(
 						tmp_tpg, i_str);
-			if (dest_node_acl) {
-				atomic_inc(&dest_node_acl->acl_pr_ref_count);
-				smp_mb__after_atomic();
-			}
+			if (dest_node_acl)
+				atomic_inc_mb(&dest_node_acl->acl_pr_ref_count);
 			spin_unlock_irq(&tmp_tpg->acl_node_lock);
 
 			if (!dest_node_acl) {
@@ -1586,8 +1567,7 @@ core_scsi3_decode_spec_i_port(
 			if (core_scsi3_nodeacl_depend_item(dest_node_acl)) {
 				pr_err("configfs_depend_item() failed"
 					" for dest_node_acl->acl_group\n");
-				atomic_dec(&dest_node_acl->acl_pr_ref_count);
-				smp_mb__after_atomic();
+				atomic_dec_mb(&dest_node_acl->acl_pr_ref_count);
 				core_scsi3_tpg_undepend_item(tmp_tpg);
 				ret = TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE;
 				goto out_unmap;
@@ -1646,8 +1626,7 @@ core_scsi3_decode_spec_i_port(
 		if (core_scsi3_lunacl_depend_item(dest_se_deve)) {
 			pr_err("core_scsi3_lunacl_depend_item()"
 					" failed\n");
-			atomic_dec(&dest_se_deve->pr_ref_count);
-			smp_mb__after_atomic();
+			atomic_dec_mb(&dest_se_deve->pr_ref_count);
 			core_scsi3_nodeacl_undepend_item(dest_node_acl);
 			core_scsi3_tpg_undepend_item(dest_tpg);
 			ret = TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE;
@@ -3167,15 +3146,13 @@ core_scsi3_emulate_pro_register_and_move(struct se_cmd *cmd, u64 res_key,
 		if (!dest_tf_ops)
 			continue;
 
-		atomic_inc(&dest_se_tpg->tpg_pr_ref_count);
-		smp_mb__after_atomic();
+		atomic_inc_mb(&dest_se_tpg->tpg_pr_ref_count);
 		spin_unlock(&dev->se_port_lock);
 
 		if (core_scsi3_tpg_depend_item(dest_se_tpg)) {
 			pr_err("core_scsi3_tpg_depend_item() failed"
 				" for dest_se_tpg\n");
-			atomic_dec(&dest_se_tpg->tpg_pr_ref_count);
-			smp_mb__after_atomic();
+			atomic_dec_mb(&dest_se_tpg->tpg_pr_ref_count);
 			ret = TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE;
 			goto out_put_pr_reg;
 		}
@@ -3271,10 +3248,8 @@ after_iport_check:
 	spin_lock_irq(&dest_se_tpg->acl_node_lock);
 	dest_node_acl = __core_tpg_get_initiator_node_acl(dest_se_tpg,
 				initiator_str);
-	if (dest_node_acl) {
-		atomic_inc(&dest_node_acl->acl_pr_ref_count);
-		smp_mb__after_atomic();
-	}
+	if (dest_node_acl)
+		atomic_inc_mb(&dest_node_acl->acl_pr_ref_count);
 	spin_unlock_irq(&dest_se_tpg->acl_node_lock);
 
 	if (!dest_node_acl) {
@@ -3288,8 +3263,7 @@ after_iport_check:
 	if (core_scsi3_nodeacl_depend_item(dest_node_acl)) {
 		pr_err("core_scsi3_nodeacl_depend_item() for"
 			" dest_node_acl\n");
-		atomic_dec(&dest_node_acl->acl_pr_ref_count);
-		smp_mb__after_atomic();
+		atomic_dec_mb(&dest_node_acl->acl_pr_ref_count);
 		dest_node_acl = NULL;
 		ret = TCM_INVALID_PARAMETER_LIST;
 		goto out;
@@ -3313,8 +3287,7 @@ after_iport_check:
 
 	if (core_scsi3_lunacl_depend_item(dest_se_deve)) {
 		pr_err("core_scsi3_lunacl_depend_item() failed\n");
-		atomic_dec(&dest_se_deve->pr_ref_count);
-		smp_mb__after_atomic();
+		atomic_dec_mb(&dest_se_deve->pr_ref_count);
 		dest_se_deve = NULL;
 		ret = TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE;
 		goto out;
@@ -3879,8 +3852,7 @@ core_scsi3_pri_read_full_status(struct se_cmd *cmd)
 		se_tpg = pr_reg->pr_reg_nacl->se_tpg;
 		add_desc_len = 0;
 
-		atomic_inc(&pr_reg->pr_res_holders);
-		smp_mb__after_atomic();
+		atomic_inc_mb(&pr_reg->pr_res_holders);
 		spin_unlock(&pr_tmpl->registration_lock);
 		/*
 		 * Determine expected length of $FABRIC_MOD specific
@@ -3893,8 +3865,7 @@ core_scsi3_pri_read_full_status(struct se_cmd *cmd)
 			pr_warn("SPC-3 PRIN READ_FULL_STATUS ran"
 				" out of buffer: %d\n", cmd->data_length);
 			spin_lock(&pr_tmpl->registration_lock);
-			atomic_dec(&pr_reg->pr_res_holders);
-			smp_mb__after_atomic();
+			atomic_dec_mb(&pr_reg->pr_res_holders);
 			break;
 		}
 		/*
@@ -3955,8 +3926,7 @@ core_scsi3_pri_read_full_status(struct se_cmd *cmd)
 				se_nacl, pr_reg, &format_code, &buf[off+4]);
 
 		spin_lock(&pr_tmpl->registration_lock);
-		atomic_dec(&pr_reg->pr_res_holders);
-		smp_mb__after_atomic();
+		atomic_dec_mb(&pr_reg->pr_res_holders);
 		/*
 		 * Set the ADDITIONAL DESCRIPTOR LENGTH
 		 */
diff --git a/drivers/target/target_core_transport.c b/drivers/target/target_core_transport.c
index 0b43761ed85f..115632ee3ec8 100644
--- a/drivers/target/target_core_transport.c
+++ b/drivers/target/target_core_transport.c
@@ -752,8 +752,7 @@ void target_qf_do_work(struct work_struct *work)
 
 	list_for_each_entry_safe(cmd, cmd_tmp, &qf_cmd_list, se_qf_node) {
 		list_del(&cmd->se_qf_node);
-		atomic_dec(&dev->dev_qf_count);
-		smp_mb__after_atomic();
+		atomic_dec_mb(&dev->dev_qf_count);
 
 		pr_debug("Processing %s cmd: %p QUEUE_FULL in work queue"
 			" context: %s\n", cmd->se_tfo->get_fabric_name(), cmd,
@@ -1721,8 +1720,7 @@ static bool target_handle_task_attr(struct se_cmd *cmd)
 			 cmd->t_task_cdb[0], cmd->se_ordered_id);
 		return false;
 	case MSG_ORDERED_TAG:
-		atomic_inc(&dev->dev_ordered_sync);
-		smp_mb__after_atomic();
+		atomic_inc_mb(&dev->dev_ordered_sync);
 
 		pr_debug("Added ORDERED for CDB: 0x%02x to ordered list, "
 			 " se_ordered_id: %u\n",
@@ -1739,8 +1737,7 @@ static bool target_handle_task_attr(struct se_cmd *cmd)
 		/*
 		 * For SIMPLE and UNTAGGED Task Attribute commands
 		 */
-		atomic_inc(&dev->simple_cmds);
-		smp_mb__after_atomic();
+		atomic_inc_mb(&dev->simple_cmds);
 		break;
 	}
 
@@ -1844,8 +1841,7 @@ static void transport_complete_task_attr(struct se_cmd *cmd)
 		return;
 
 	if (cmd->sam_task_attr == MSG_SIMPLE_TAG) {
-		atomic_dec(&dev->simple_cmds);
-		smp_mb__after_atomic();
+		atomic_dec_mb(&dev->simple_cmds);
 		dev->dev_cur_ordered_id++;
 		pr_debug("Incremented dev->dev_cur_ordered_id: %u for"
 			" SIMPLE: %u\n", dev->dev_cur_ordered_id,
@@ -1856,8 +1852,7 @@ static void transport_complete_task_attr(struct se_cmd *cmd)
 			" HEAD_OF_QUEUE: %u\n", dev->dev_cur_ordered_id,
 			cmd->se_ordered_id);
 	} else if (cmd->sam_task_attr == MSG_ORDERED_TAG) {
-		atomic_dec(&dev->dev_ordered_sync);
-		smp_mb__after_atomic();
+		atomic_dec_mb(&dev->dev_ordered_sync);
 
 		dev->dev_cur_ordered_id++;
 		pr_debug("Incremented dev_cur_ordered_id: %u for ORDERED:"
@@ -1915,8 +1910,7 @@ static void transport_handle_queue_full(
 {
 	spin_lock_irq(&dev->qf_cmd_lock);
 	list_add_tail(&cmd->se_qf_node, &cmd->se_dev->qf_cmd_list);
-	atomic_inc(&dev->dev_qf_count);
-	smp_mb__after_atomic();
+	atomic_inc_mb(&dev->dev_qf_count);
 	spin_unlock_irq(&cmd->se_dev->qf_cmd_lock);
 
 	schedule_work(&cmd->se_dev->qf_work_queue);
diff --git a/drivers/target/target_core_ua.c b/drivers/target/target_core_ua.c
index 101858e245b3..1738b1646988 100644
--- a/drivers/target/target_core_ua.c
+++ b/drivers/target/target_core_ua.c
@@ -161,8 +161,7 @@ int core_scsi3_ua_allocate(
 		spin_unlock(&deve->ua_lock);
 		spin_unlock_irq(&nacl->device_list_lock);
 
-		atomic_inc(&deve->ua_count);
-		smp_mb__after_atomic();
+		atomic_inc_mb(&deve->ua_count);
 		return 0;
 	}
 	list_add_tail(&ua->ua_nacl_list, &deve->ua_list);
@@ -174,8 +173,7 @@ int core_scsi3_ua_allocate(
 		nacl->se_tpg->se_tpg_tfo->get_fabric_name(), unpacked_lun,
 		asc, ascq);
 
-	atomic_inc(&deve->ua_count);
-	smp_mb__after_atomic();
+	atomic_inc_mb(&deve->ua_count);
 	return 0;
 }
 
@@ -189,8 +187,7 @@ void core_scsi3_ua_release_all(
 		list_del(&ua->ua_nacl_list);
 		kmem_cache_free(se_ua_cache, ua);
 
-		atomic_dec(&deve->ua_count);
-		smp_mb__after_atomic();
+		atomic_dec_mb(&deve->ua_count);
 	}
 	spin_unlock(&deve->ua_lock);
 }
@@ -250,8 +247,7 @@ void core_scsi3_ua_for_check_condition(
 		list_del(&ua->ua_nacl_list);
 		kmem_cache_free(se_ua_cache, ua);
 
-		atomic_dec(&deve->ua_count);
-		smp_mb__after_atomic();
+		atomic_dec_mb(&deve->ua_count);
 	}
 	spin_unlock(&deve->ua_lock);
 	spin_unlock_irq(&nacl->device_list_lock);
@@ -309,8 +305,7 @@ int core_scsi3_ua_clear_for_request_sense(
 		list_del(&ua->ua_nacl_list);
 		kmem_cache_free(se_ua_cache, ua);
 
-		atomic_dec(&deve->ua_count);
-		smp_mb__after_atomic();
+		atomic_dec_mb(&deve->ua_count);
 	}
 	spin_unlock(&deve->ua_lock);
 	spin_unlock_irq(&nacl->device_list_lock);